|Publication number||US4354164 A|
|Application number||US 06/079,671|
|Publication date||Oct 12, 1982|
|Filing date||Sep 27, 1979|
|Priority date||Sep 27, 1979|
|Publication number||06079671, 079671, US 4354164 A, US 4354164A, US-A-4354164, US4354164 A, US4354164A|
|Inventors||Shanti S. Gupta|
|Original Assignee||Communications Satellite Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (4), Referenced by (7), Classifications (10), Legal Events (2)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention is related to my co-pending application entitled, "High Resolution Frequency Synthesizer", Ser. No. 79,603, filed on Sept. 27, 1979. In particular, this invention is directed to a phase lock loop applicable to the frequency synthesis techniques disclosed in that co-pending application. In my co-pending patent application, a system is disclosed having a center frequency and a programmable finite pull-in and holding range. The pull-in and holding range is the same and requires no filters.
In conventional phase lock loop circuitry (PLL), pull-in and holding ranges are different. Therefore, such circuitry generally requires the use of filters within the loop. The present invention removes such a requirement.
In the co-pending application, the reference oscillator and clock filter are not affected in the circuitry. Accordingly, there are multiple phase lock loops on the same reference clock without affecting each other and they may be maintained in an edge synchronous manner to the reference clock frequency. Accordingly, the jitter and the output frequency remain the same over the entire PLL range. It is of course a function of the reference clock frequency. This is contrasted with standard PLL technology where jitter becomes extremely large with large frequency division.
As a result, resolution can be controlled within very narrow limits even when the ratio of the output frequency to the synchronization frequency is at extremes, whether they be large or small. Moreover, a known time base repetitive pulse may be used for alignment of frequency even when the ratio of frequency to time base pulse is exceptionally large. Such technique is useful in time division multiple access (TDMA) applications where a large frame length is used and a large number of frequencies must be adjusted due to doppler effect of satellite movement and the stability of the reference frequency.
This invention pertains specifically to the phase lock loop aspects of such frequency synthesizers and is based on the concept that when different frequencies are counted to a fixed number, on a repetitive basis, they will generate a different time base. Conversely, the number of bits over a fixed time base is different for different frequencies. These two principles are used in this invention for the development of the phase lock loop. The error is detected at the end of the count and a programmed frequency change then takes place if there is an error. Otherwise, the synthesizer is kept at the programmed center frequency. This technique of error detection is used at the end of each count.
This invention will be described in greater detail with respect to the accompanying drawings and the description of the preferred embodiment which follow.
FIG. 1 is a waveform diagram showing a first case where two frequencies are available: one, a master clock frequency and the other a slave frequency;
FIG. 2 is a block diagram showing details of the frequency change monitor and relates this invention to the FIG. 1 timing diagram;
FIG. 3 is a block diagram showing a phase locked loop in accordance with this invention;
FIG. 4 is a timing diagram showing a second case where a synchronizing pulse appears after a fixed interval; and
FIG. 5 is a block diagram showing a second embodiment of the phase lock loop in accordance with the present invention.
Referring now to FIGS. 1 and 2, a first case is shown where two frequencies are available, one being a master clock frequency and the other being a slave or output frequency. As shown in FIG. 1, the master frequency is shown as curve (a) and the slave frequency is shown in three cases as (b), (c) and (d). In particular, in the second curve (b), the slave frequency is a higher frequency than that of the master frequency and in the second case (c) the slave frequency is a lower frequency than that of the master frequency. Finally, in the waveform (d) the slave frequency and the master frequency are the same. Accordingly, all three cases are covered vis-a-vis frequency variation in the first case. As shown in FIG. 1, the frequency different between the curves (a) and (b) is α, and in the case between the master and the lower frequency (c), the difference is shown as β.
The slave frequency is generated utilizing digital synthesizer techniques as disclosed in my co-pending application. The two frequencies are applied to two different counters which are each set to count to a number N.
The Nth bit of the desired frequency is used for a loading that is synchronization of the slave frequency counter. Using the principles of this invention, of the generation of different time bases by two different frequencies, the Nth bit of the master and slave frequencies are compared to generate the correction program for feedback change in frequency. This correction program feedback will change the number Nk as described in my co-pending application. To the extent that the disclosure on the derivation of Nk is needed, the following discussion is presented to explain the function of Nk in the context of the frequency synthesizer of my co-pending application. Assume that the acceptable jitter in the TDMA TIM clock frequency is A nanosecond. The lowest reference clock frequency cy fS is:
fS =1/A MHz (1)
Let the desired TDMA TIM frequency be fD.
Ratio of two frequencies N=fS /fD (2)
N=NE +NR (3)
NE =Even number, preferably highest, but cannot be more than N, and
NR =Remainder; can be whole or fraction, but cannot be negative.
From equations (2) and (3),
fS /fD =NE +NR (4)
fS =NE fD +NR fD (5)
NE fD =fx (6)
NR fD =fS -fx
fx =frequency or number of bits per second which is an even multiple of desired frequency, and
NR fD =Difference of bits per second between master and fx.
Divide equation (5) by NR fD ##EQU1## Let NKH1 =first higher whole number to NK,
NKL1 =first lower whole number to NK,
NKH2 =(NKH1 +1),
NKL2 =(NKL1 -1)
and so on.
QH1 =multiplier of NKH1
QH2 =multiplier of NKH2
QL1 =multiplier of NKL1
QL2 =multiplier of NKL2
and so on. Let ##EQU2##
FIG. 2 shows the frequency change monitor 24. In the Figure, two flip flops and a standard NOR circuit are shown. These components generate three logic correction signals, as will be discussed later, which provide control information to divider 12, shown in FIG. 3. As noted in FIG. 2 which shows the frequency change monitor, three outputs would be obtained, a higher frequency, a lower frequency and the same frequency. These three outputs are then fit into a program to change or keep the same value of Nk for the frequency. Therefore, a block diagram of the phase lock loop in accordance with this invention is shown in FIG. 3.
FIG. 3 shows the phase lock loop circuit aspects of this invention in the dotted box 10. The remaining portions of the block diagram in FIG. 3 comprise the frequency synthesizer as described in my co-pending application. That is, as described in that application, a reference oscillator supplies the signal at frequency fS to the divider 12 for division by Nk. The reference frequency and the divider signal are then supplied to the AND gate 14 which provides an output fx which is fed to a second divider 16 for division by the value NE divided by 2. Finally, to obtain a symmetrical waveform, the resulting signal is divided by 2 in divider 8 to produce the final desired signal FD. In accordance with this invention, a phase lock loop 10 is used to define a program change for the value of Nk. The master frequency in the form of clock pulses is provided to the counter 20 for division by N. The counter is loaded with the Nth bit as shown to count to N. The output of counter 20 is fed to gate 22 and a frequency change monitor 24 the details of which are shown in FIG. 2. Gate 22 is used to provide a load signal for a second counter 26, supplying a feedback signal to the gate 22 and the frequency change monitor 24.
As a result, as shown in FIG. 3, the master frequency is counted to the set count N in counter 20 providing a first input to the frequency change monitor 24 in the form of the Nth bit of the master frequency. The Nth bit of the slave frequency provides a second input to the frequency change monitor along line 28. This Nth bit is also supplied as an output to the OR gate 22 which is used to load the counter 26. The counter 26 receives the synthesized output fD as the feedback input to the loop.
Consequently, the frequency change monitor receives the Nth bits of both the master and slave frequencies to compare them and generate the correction program for the feedback change in frequency to the divider Nk. The frequency change monitor 24 essentially comprises 2 flip-flops having a synchronizing pulse input coupled to the cl inputs such as in a MC10131 f-f which is depicted at page 3-35 in the 1978 MOTOROLA MECL Integrated Circuits handbook. The second input from the divider 26 is supplied to the SD terminals of such f-f elements.
The outputs of the two flip-flops are used in combination with the output of a standard NOR gate to provide three logic correction signals to control the value Nk of divider 12. A first control signal taken from the output of the first flip-flop will be high if the synchronizing pulse is at a higher frequency than the Nth pulse. Conversely, this control signal will be low where the synchronizing pulse is lower than, or at the same frequency as the Nth pulse. A second control signal is taken from the output of the NOR gate and will be high when the synchronizing pulse is the same frequency as the Nth pulse and low otherwise. A third control signal is taken from the output of the second flip-flop and will be high when the synchronizing pulse is lower than the frequency of the Nth pulse and will be low when the synchronizing pulse is higher than, or the same as the frequency of the Nth pulse.
Referring now to FIGS. 4 and 5, a second embodiment of this invention is shown in which the synchronizing pulse appears after a fixed time interval, that is after the time base. The time base is used for generating frequency in accordance with the formula
frequency (f)=no. of bits in time base (N)/time base (T)
This can also be expressed as
The program synthesized frequency, that is, the slave frequency, can then be applied to a counter to count up to N. This counter is synchronized by the fixed time base pulse. The Nth bit is divided into three components, as shown in FIG. 4. These components are positioned relative to the synchronizing pulses which are shown in the bottom of the figure. The first is the component dealing with the first half of the Nth pulse from the leading edge to the center of the synchronizing pulse. The second component would be the second half of the Nth pulse from the edge of the synchronizing pulse to the trailing edge of the Nth pulse. Finally, at the exact center frequency about the synchronizing pulse, the third component is shown.
Accordingly, if the synchronizing pulse appears in the shaded area shown in waveform (d) of FIG. 4, then it is apparent that no change is needed since the Nth bit is centered about the synchronizing pulse. If the synchronizing pulse appears in the shaded area of waveform (b), but not in the portion of the shaded area of (d), the slave frequency is lower. Conversely, if the synchronizing pulse appears in the shaded area of waveform (c) of FIG. 4, but not in the shaded area of waveform (d), the slave frequency is higher.
Given this status, the second embodiment of FIG. 5 can be shown.
The elements of FIG. 5, which are the same as those in FIG. 3, have been consistently labeled and will not be discussed in detail. As shown, in FIG. 5, the error detector and program changer is used to compare the synchronizing pulse with the Nth pulse delivered from divider 20. Based on the discussion of the timing chart in FIG. 4, a correction to Nk is made, depending on the relationship between the frequency of the component of the Nth pulse and the synchronizing pulse. As in the prior example, the Nth pulse is used to load the counter 20 such that it counts to a specific fixed number. Also, it should be noted that the output in the form of the symmetrical desired output frequency fD is used to provide the input to the counter 20 to close the loop.
As seen, this technique is useful where every fraction of a frequency can be adjusted and the output frequency will be edge synchronous with the incoming frequency. In the case of repetitive time base pulses, such as a satellite frame pulse of variable width, the TIM clock frequency will lock onto it. Therefore, satellite doppler and reference oscillator stability effects are corrected in this system to maintain synchronous operation.
It is apparent that modifications of this system can be achieved without departing from the essential scope of the invention.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3731219 *||Jun 13, 1972||May 1, 1973||Int Standard Electric Corp||Phase locked loop|
|US3983498 *||Nov 13, 1975||Sep 28, 1976||Motorola, Inc.||Digital phase lock loop|
|US4024460 *||Jul 7, 1975||May 17, 1977||Hewlett-Packard Company||Electronic line stretcher|
|US4105947 *||Sep 16, 1977||Aug 8, 1978||Rca Corporation||Pulse wave phase and frequency detector|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US4558360 *||Oct 3, 1983||Dec 10, 1985||Zenith Electronics Corporation||Decoder lockout defeat circuit|
|US4802235 *||Apr 26, 1985||Jan 31, 1989||Comven, Inc.||Subscriber unit for a flexible communication system|
|US4947382 *||Apr 11, 1989||Aug 7, 1990||Vista Labs, Inc.||Direct digital locked loop|
|US7567108||Jun 13, 2007||Jul 28, 2009||Realtek Semiconductor Corp.||Apparatus and method for generating clock signal|
|US7990198||Jun 22, 2009||Aug 2, 2011||Realtek Semiconductor Corp.||Apparatus and method for generating clock signal|
|US20050156649 *||Jan 13, 2005||Jul 21, 2005||Sen-Huang Tang||Apparatus and method for generating clock signal|
|US20070247207 *||Jun 13, 2007||Oct 25, 2007||Sen-Huang Tang||Apparatus and method for generating clock signal|
|U.S. Classification||331/1.00A, 331/14, 327/7, 327/12, 331/16, 331/25, 327/231|
|Feb 15, 1983||CC||Certificate of correction|
|Oct 8, 1993||AS||Assignment|
Owner name: COMSAT CORPORATION, MARYLAND
Free format text: CHANGE OF NAME;ASSIGNOR:COMMUNICATIONS SATELLITE CORPORATION;REEL/FRAME:006711/0455
Effective date: 19930524